Diagnose

This is essentially a linter for your AI workflows. It scores your setup across five dimensions: prompt quality, context efficiency, tool health, architecture fitness, and safety. Each dimension gets a 1-5 rating with specific findings, then it maps the gaps to concrete Maestro commands like /fortify or /streamline. The scoring rubric is strict, anything below a 3 means you've got real problems to fix. What I like here is that it doesn't just tell you what's wrong, it tells you which command to run next. The tool health dimension is especially practical, flagging things like too many tools (over 7 is a red flag) or missing error handling. Good for periodic health checks or when something feels off but you can't pinpoint why.

Install to Claude Code

npx -y skills add sharpdeveye/maestro --skill diagnose --agent claude-code